/*
 * crypto.c
 * (c)Steve Parker, 31st Oct 2003
 * GPL v2 for now.
 * http://www.gnu.org/copyleft/gpl.html
 *
 * Aimed at challenges such as 
 * http://www.thawte.com/cryptochallenge/lunchHour.html
 * Goals: 
 *  Identify letters by frequency analysis
 *  Confirm by digrams
 *  Confirm agian by Trigraphs
 * Resources:
 *   http://library.thinkquest.org/28005/flashed/thelab/cryptograms/frequency.shtml
*/

/* 
Bmxbzs cfbs jo njoe uibu zpvs pxo sftpmvujpo up tvddffe jt npsf jnqpsubou
uibo boz pof uijoh.
Becomes: (ie, out-by-one)
Always bear in mind that your own resolution to succeed is more important than any one thing. Abraham Lincoln 
*/




#include <stdio.h>
#define MAXCRYPTO 999999
	// Longest string to be translated

// For now, assume only spaces and a-z (not even A-Z)
// For testing purposes, I'm taking plain text input and hoping to re-create plain-text output.
// That's the basic test for frequency analysis, after all!
// With more advanced checks, such details may not be as necessary (eg, "th" go together, etc)
//, '\'', ',', ',', '-','1','2','3','4','5','6','7','8','9','0'};

// http://library.thinkquest.org/28005/flashed/thelab/cryptograms/frequency.shtml
//char english[]={ ' ', 'e', 't', 'n', 'r', 'i', 'o', 'a', 's', 'd', 'h', 'l', 'c', 'f', 'p', 'u', 'm', 'y', 'g', 'w', 'v', 'b', 'x', 'k', 'q', 'j', 'z' };

// http://paul.oniony.com/sections/information/frequencyanalysis.html
//char english[]={' ', 'e', 't', 'a', 'o', 'n', 'i', 'r', 's', 'h', 'd', 'l', 'c', 'u', 'm', 'f', 'p', 'y', 'g', 'w', 'b', 'v', 'k', 'x', 'j', 'q', 'z' };

// From Act 1, Scene 1, "All's Well that ends well" (http://the-tech.mit.edu/Shakespeare/)
// Simple version:
//char english[]={ ' ', 'e', 't', 'i', 'a', 'o', 's', 'n', 'r',  'h', 'l', 'd', 'u', 'm', 'w', 'y', 'f', 'c', 'g', 'b', 'p', 'v', 'k', 'x', 'q', 'j', 'z' };
// Full version:
//char english[] = { ' ', 'e', 't', 'i', 'o', 'a', 's', 'n', 'r', 'h', 'l', 'd', 'u', 'm', 'w', 'y', ',', 'f', 'c', 'g', '*', 'b', 'E', 'p', 'v', '.', 'k', 'A', 'L', 'T', '\'', ';', 'S', 'I', 'N', 'H', 'O', 'R', ':', 'U', 'P', 'W', 'M', 'B', '/', '?', 'F', '!', 'C', '-', 'x', 'q', 'j', 'Y', 'G', 'V', 'K', '[', ']', 'D', 'z' }; 

// From http://www.thingsmygirlfriendandihavearguedabout.com/ :
char english[]={ ' ', 'e', 't', 'a', 'o', 'n', 'i', 's', 'r', 'h', 'l', 'd', 'u', 'g', 'm', 'c', 'y', 'w', 'f', '\'', 'p', ',', '.', 'b', 'v', 'I', 'k', '/', '-', 'M', 'T', '?', 'S', 'A', 'W', 'x', 'j', ':', 'O', 'B', '*', 'H', 'N', 'Y', ')', 'D', 'E', '!', 'G', '(', 'C', 'q', 'F', 'L', 'P', 'z', '"', 'R', 'K', 'V', ';', 'J', '1', '<', '>', 'U', '2', '0', '3', '8', '9', '[', ']', '5', '4', 'Q', '_', '7', '=', '+', 'Z', '~', '%', '6', '#', '&', '@', 'X' }; 

// From freq_analysis.c:
//char english[]={ ' ', 'i', 't', 'r', 'f', ';', 'n', 'c', 'e', 'a', '(', ')', 's', 'o', '[', ']', '.', 'p', 'l', '=', 'h', '+', '*', '0', ',', 'u', '{', '}', '5', '/', 'g', '2', '"', 'b', 'd', '1', '<', 'j', 'm', 'q', 'w', '\'', 'v', '9', 'y', 'P', '%', 'A', 'C', 'O', 'T', '\\', '-', '3', 'I', 'L', 'M', 'R', 'S', 'X', 'Y', '_', 'k', ' ', '#', '&', ':', '>', '!', '4', '6', '7', '8', 'E', 'F', 'G', 'x' };

struct freq
{
  int f;
  char c;
  char t;
};

void
bubblesort (struct freq a[], int n)
{
  int i, j;
  struct freq tmp;

  for (j = 0; j < n - 1; j++)
    {
      for (i = 0; i < n - 1; i++)
	{
	  if (a[i].f < a[i + 1].f)
	    {
	      tmp = a[i];
	      a[i] = a[i + 1];
	      a[i + 1] = tmp;
	    }
	}
    }
}

void
freq_analysis (char *p)
{
  int i, j;
  int k;
  struct freq f[255];
  char p2[MAXCRYPTO];

  for (i = 0; i < 255; i++)
    {
      f[i].f = 0;
      f[i].c = '\0';
      f[i].t = '\0';
    }
  for (i = 0; i < strlen (p); i++)
    {
      if (p[i] > 31)
	{
	  f[(int) p[i]].f++;
	  f[(int) p[i]].c = p[i];
	}
    }

  bubblesort (f, 256);

  j = 0;
  for (i = 0; i < 255; i++)
    {
      if (f[i].f > 0)
	{
	  while (f[i].c < 0)
	    f[i].c += 255;
	  if (((int) f[i].c < 127) && ((int) f[i].c != 10))
	    {
	      printf ("%4d : %c (%03d) = %c (%03d)\n", f[i].f, f[i].c, f[i].c,
		      english[j], english[j]);
	      for (k = 0; k < MAXCRYPTO; k++)
		{
		  if (p[k] == f[i].c)
		    {
		      p2[k] = english[j];
		    }
		  else
		    {
		      if (p[k] < 32)
			p2[k] = p[k];
		    }
		}
	      j++;
	    }
	}
    }
  printf ("\nTranslation:\n");
  //printf("%s\n", p);
  //printf("Becomes:\n");
  p2[strlen (p) - 2] = '\0';
  printf ("%s\n", p2);
}

int
main (int argc, char *argv[])
{

//  char *p="Js Jr qntrjamd un gzjk jm nzox xzzr...xgjkf sp rvbddfc jr qntrjamd pmmx jm pmf vbx.";
  char *p;
  char *s;
  FILE *in;

  p = (char *) malloc (MAXCRYPTO);
  s = (char *) malloc (255);

  in = fopen (argv[1], "rb");
  while (fgets (s, 255, in))
    {
      strcat (p, s);
    }
  fclose (in);
  //printf("Code: %s\n", p);

  freq_analysis (p);
  free (p);

  return 0;
}